Systemic rheumatic diseases have been shown to produce antinuclear antibodies in many previous studies. This property has proved useful both for diagnostic purposes and for the identification and characterisation of novel nuclear antigens. In a blind study, a panel of 24 sera from patients with a wide range of systemic rheumatic disease was assayed for autoantibodies specific for the nucleus by whole cell immunofluorescence and Western blotting. Some well characterised clinical conditions were identified by the specific immunofluorescence staining patterns observed, and by the molecular weights of the polypeptides reacting on blots. Apparently novel staining patterns in cells and on blots were also observed and two sera in particular were chosen for further study. The autoantibodies present in these sera and in sera from additional patients with the same conditions allowed the identification of cDNA clones from a human expression library. A group of sera from scleroderma patients which recognised a polypeptide of 95kDa led to the isolation of 3 clones which cross-hybridised with each other, and when sequenced proved to code for DNA topoisomerase I. This protein has been shown to be identical to the marker antigen Scl-70 originally estimated to have a molecular weight of 70kDa (Shero et al, 1986). The second group of sera from MCTD patients recognising a polypeptide of 38kDa on western blots led to the isolation of a single human cDNA showing im75% nucleic acid sequence identity and im85% amino acid sequence identity with the m9 transcript of the Drosophila development gene Enhancer of Split E(spl). E(spl) m9 forms part of a complex locus in Drosophila which interacts with other neurogenic genes during the initial stages of neurogenesis. Loss or a mutation in any one of those genes results in neural hyperplasis due to incorrect determination of cell fate. However this protein is not the 38kDa antigen detected on Western blots, as the cDNA is incorporated in the reverse orientation relative to the direction of transcription within the λ vector. The isolation of the human homologue of E(spl) m9 clone was probably due to the selection of a spurious negative clone, which had not been eliminated in the rounds of screening and which remained on the final plate.